Autonomous inspection task allocation optimization method for low-altitude flight unmanned aerial vehicle

Abstract

The invention discloses an autonomous inspection task allocation optimization method for a low-altitude flight unmanned aerial vehicle, and belongs to the technical field of intelligent control of unmanned aerial vehicles. The method comprises the following steps: acquiring task attributes, an unmanned aerial vehicle state, a geographical environment and real-time meteorological data through a data acquisition module; the environment modeling module is fused with geographic and meteorological data to construct a three-dimensional dynamic environment model; the dynamic allocation module takes the minimization of the number of times of operation of the unmanned aerial vehicle as a target, combines the single maximum flight distance of the unmanned aerial vehicle and path safety threshold constraints, adopts an improved saving algorithm to cluster to perform task grouping, optimizes a task sequence based on a vehicle path problem model with a time window, and generates a task allocation scheme; and the control terminal issues a task allocation scheme to the unmanned aerial vehicle, the unmanned aerial vehicle executes the scheme and adjusts a path according to real-time meteorological changes, and a multi-dimensional inspection report is generated after the task is completed. According to the invention, the task allocation efficiency and the flight safety in a complex environment are improved, and the inspection cost is reduced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent control of unmanned aerial vehicles, and specifically discloses an autonomous inspection task allocation optimization method for low-altitude flying unmanned aerial vehicles. BACKGROUND

[0002] Unmanned aerial vehicles have been widely used in multiple fields due to their flexibility, efficiency and intelligence, significantly improving work efficiency and reducing costs. However, the current unmanned aerial vehicle inspection task allocation method has the following limitations: ‌Insufficient environmental adaptability‌: Traditional methods are mostly based on static geographic data to plan paths, without integrating real-time weather such as wind speed and turbulence, which dynamically affect flight safety and energy consumption, resulting in a high risk of planned paths failing in actual flight.

[0003] ‌Task allocation efficiency needs to be improved‌: Task grouping mostly relies on manual experience or simple distance clustering, without considering task priority, time window constraints and unmanned aerial vehicle endurance restrictions, which can easily cause flight redundancy or task overtime.

[0004] ‌Insufficient safety quantification‌: Risk factors such as flight restricted areas and obstacles are usually judged in a binary manner, lacking detailed safety assessment of complex low-altitude environments such as building group wind fields and sudden weather changes.

[0005] ‌Lack of dynamic response‌: Path adjustment during flight relies on manual intervention and cannot autonomously avoid sudden strong winds, sudden visibility drops and other weather threats, affecting task success rate.

[0006] Therefore, it is necessary to invent an autonomous inspection task allocation optimization method for low-altitude flying unmanned aerial vehicles to solve the above problems. SUMMARY

[0007] In order to overcome the above-mentioned technical defects of the prior art, the present application provides an autonomous inspection task allocation optimization method for low-altitude flying unmanned aerial vehicles, which acquires task attributes, unmanned aerial vehicle states, geographic environment and real-time weather data through a data acquisition module; an environment modeling module fuses geographic and weather data to construct a three-dimensional dynamic environment model; a dynamic allocation module minimizes the number of unmanned aerial vehicle operations, combines the maximum single flight distance of the unmanned aerial vehicle and the path safety threshold constraint, uses an improved saving algorithm for clustering to group tasks, optimizes the task sequence based on a vehicle routing problem model with a time window, and generates a task allocation scheme; a control terminal issues the task allocation scheme to the unmanned aerial vehicle, the unmanned aerial vehicle executes the scheme and adjusts the path according to real-time weather changes, and generates a multi-dimensional inspection report after the task is completed, effectively solving the problems mentioned in the background art.

[0008] To achieve the above object, the application provides the following technical scheme: a low-altitude flight unmanned aerial vehicle autonomous inspection task allocation optimization method, characterized by comprising a data acquisition module, an environment modeling module, a dynamic allocation module, an unmanned aerial vehicle and a control terminal, and specifically comprising the following steps: S1, acquiring task attribute data, unmanned aerial vehicle state data, geographical environment data and real-time weather data through the data acquisition module; S2, the environment modeling module fuses the geographical environment data and the real-time weather data to construct a three-dimensional dynamic environment model; S3, the dynamic allocation module generates a task allocation scheme based on the unmanned aerial vehicle single maximum flight distance and the minimum path safety threshold constraint, with the minimum operation number as the target; S4, the control terminal issues a task to the unmanned aerial vehicle according to the task allocation scheme; S5, the unmanned aerial vehicle executes the task and adjusts the flight path according to the real-time weather data, and the control terminal generates an inspection report after the task is completed; On the basis of the above embodiment, the task attribute data includes task location coordinates, task priority and window time; the unmanned aerial vehicle state data includes the remaining endurance mileage of the unmanned aerial vehicle; the geographical environment data includes digital elevation model, obstacle distribution data, no-fly zone boundary coordinates and land cover type data; the real-time weather data includes wind speed, wind direction, temperature, precipitation probability, atmospheric visibility and turbulence intensity index.

[0009] On the basis of the above embodiment, the construction process of the three-dimensional dynamic environment model comprises: The digital elevation model and the obstacle distribution data are superimposed to generate a static terrain skeleton; The real-time weather data is interpolated by a finite element grid to generate a wind speed and turbulence intensity spatial distribution heat map; The no-fly zone boundary coordinates and the land cover type data are fused to mark the safety level of the flyable area.

[0010] On the basis of the above embodiment, the safety level calculation formula is: In the formula, S is the safety level score, V is the real-time wind speed, T i is the turbulence intensity index, C t is the land cover type coefficient, w1, w2 and w3 are weight coefficients, and k is the wind speed attenuation factor.

[0011] On the basis of the above embodiment, the task allocation scheme is generated, and the specific analysis method is: Based on the task priority and the task time window, all tasks are sorted to generate an initial task sequence; Under the constraint of the maximum flight distance of a single flight of the UAV, the initial task sequence is clustered and grouped into several subsets of tasks that can be completed in a single flight by using the improved saving algorithm, and the goal is to minimize the total flight times of the required UAVs; For each subset of tasks, a vehicle routing problem model with time windows is constructed based on the current location of the UAV, the task location coordinates, the estimated flight time of the inspection path, and the time window constraints. The genetic algorithm is used to solve the model to optimize the execution order of the tasks in the group, and the goal is to minimize the total flight distance and time of the flight and ensure that all tasks are completed within their time windows; According to the safety level of the flyable area, the task safety index is calculated, and it is checked whether there is a time window conflict, the task safety index is lower than the threshold, or the single maximum flight distance constraint is exceeded in the optimized task sequence. If there is a conflict, adjust the task grouping or sequence order until all constraints are met; The task allocation scheme is output, which includes the number of UAVs, the task sequence executed by each flight, the estimated take-off time of each flight, the total flight distance, and the estimated execution time of each task.

[0012] On the basis of the above embodiment, the calculation method of the task safety index is: In the formula, I is the task safety index, n is the number of paths covered in the task subset, S i is the safety level score of the ith path, l i is the length of the ith path, L is the total length of the n paths, P is the task priority, β is the priority weight coefficient, △T i is the turbulence intensity change rate, γ is the meteorological mutation penalty coefficient, and δ(t) is the time urgency factor.

[0013] On the basis of the above embodiment, the flight path adjustment is specifically analyzed as follows: If any of the following conditions is detected, the path dynamic adjustment is triggered: The turbulence intensity index of the current path segment exceeds the turbulence intensity threshold; The crosswind speed exceeds 70% of the upper limit of the UAV's wind resistance capacity; The atmospheric visibility decreases, resulting in a target observation clarity lower than the preset value; The adjustment strategy includes: Based on the updated environment model, the local path is replanned to avoid high turbulence areas and strong wind areas; According to the real-time visibility, the flight altitude is dynamically adjusted to ensure the effectiveness of observation; If the energy consumption of the adjusted path exceeds the remaining range, the control terminal is requested to request task termination or return instruction.

[0014] On the basis of the above-mentioned embodiments, the control terminal generates an inspection report after the task is completed, and specifically includes: Task execution timing log; Deviation analysis of actual flight path and preset path; Influence statistics of meteorological data on flight energy consumption; Path dynamic adjustment times and reason classification; Task completion state marking.

[0015] The technical effects and advantages of the present application are: 1. Fusing digital elevation model, obstacle distribution and real-time meteorological data, constructing a three-dimensional dynamic environment model containing wind speed field, turbulence intensity thermal map and safety level, providing real-time risk quantification basis for path planning; 2. Taking the minimum flight as the goal, generating a task subset through improved economy algorithm clustering, combining with genetic algorithm to optimize the task sequence in the group, and ensuring efficient execution of high-priority tasks under the constraints of time window and endurance distance; and 3. Automatically detecting turbulence threshold, crosswind threshold or insufficient visibility during flight, triggering local path re-planning, height adjustment or return request, and improving the robustness of the task under sudden weather; 4. The control terminal generates an inspection report containing deviation analysis, meteorological influence statistics, and adjustment reason classification, supporting task review and algorithm iteration optimization. BRIEF DESCRIPTION OF DRAWINGS

[0016] The present application is further illustrated by the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present application. For ordinary skilled in the art, other drawings can be obtained without creative labor on the basis of the following drawings.

[0017] Figure 1 The overall structure of the present application is shown in the figure.

[0018] Figure 2 The overall steps of the present application are shown in the figure. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0020] The present application provides a low-altitude flight unmanned aerial vehicle autonomous inspection task allocation optimization method as shown in Figure 1 The structure is as shown in Figure 1As shown, comprising a data acquisition module, an environment modeling module, a dynamic allocation module, a UAV and a control terminal; As shown in the drawings, Figure 2 As shown, the low-altitude flight unmanned aerial vehicle autonomous inspection task allocation optimization method specifically comprises the following steps: S1, acquiring task attribute data, unmanned aerial vehicle state data, geographic environment data and real-time weather data through the data acquisition module; Further, in the above technical solution, the task attribute data includes: task location coordinates, task priority and window time; the unmanned aerial vehicle state data includes: unmanned aerial vehicle remaining range; the geographic environment data includes: digital elevation model, obstacle distribution data, no-fly zone boundary coordinates, land cover type data; the real-time weather data includes: wind speed, wind direction, temperature, precipitation probability, atmospheric visibility, turbulence intensity index.

[0021] S2, the environment modeling module fuses the geographic environment data and the real-time weather data to construct a three-dimensional dynamic environment model; Further, in the above technical solution, the construction process of the three-dimensional dynamic environment model comprises: Superimposing the digital elevation model and the obstacle distribution data to generate a static terrain skeleton; Interpolating the real-time weather data through a finite element grid to generate a wind speed, turbulence intensity spatial distribution heat map; Fusing the no-fly zone boundary coordinates and the land cover type data to mark the safety level of the flyable area.

[0022] Further, in the above technical solution, the safety level calculation formula is: In the formula, S is the safety level score, V is the real-time wind speed, T i is the turbulence intensity index, C t is the land cover type coefficient, w1, w2, w3 are weight coefficients, and k is the wind speed attenuation factor.

[0023] It should be further explained that the safety level score has a value range of S∈[0, 1], and the larger the value, the safer; the land cover type coefficient C t is valued according to the inspection area type, in a preferred technical solution, when the inspection area is a no-fly zone, C t =0, when it is a building area, C t =0.2, when it is a water area, C t =0.5, when it is a vegetation area, C t =0.8, and when it is an open land, C t=1.0; the weight coefficients satisfy w1+w2+w3=1, the default values are w1=0.4, w2=0.3, and w3=0.3, and the wind speed attenuation factor has a default value k=0.2; Safety score calculation example: Scenario description: the real-time wind speed V=8 m / s in a certain area, the turbulence intensity index T i =0.7, the ground surface coverage type is a mixed area, containing 30% of the building area and 70% of the water area, the weight coefficients are w1=0.4, w2=0.3, and w3=0.3, and the wind speed attenuation factor is k=0.2; Calculation process: C t =(0.3x0.2)+(0.7x0.5)=0.06+0.35=0.41;

[0024] S=0.4x e (-0.2×8) +0.3x(1-0.7)+0.3x0.41≈0.29; Conclusion: the safety score of the area is 0.29.

[0025] S3, the dynamic allocation module generates a task allocation scheme based on the maximum single flight distance of the unmanned aerial vehicle and the minimum path safety threshold constraint, with the goal of minimizing the number of flights; Further, in the above technical solution, the task allocation scheme is generated in the following specific analysis manner: Sort all tasks based on task priority and task time window to generate an initial task sequence; It should be further pointed out that the specific manner of sorting is: first, sort the tasks according to the task priority from high to low, and sort the tasks with the same priority according to the task time window, and sort the tasks with more urgent time earlier; Taking the maximum single flight distance of the unmanned aerial vehicle as the constraint condition, an improved saving algorithm is used to cluster and group the tasks in the initial task sequence to form several groups of task subsets that can be completed in a single flight, with the goal of minimizing the total flight times of the required unmanned aerial vehicles; It should be further pointed out that the maximum single flight distance D max is 90% of the remaining endurance mileage of the unmanned aerial vehicle; The specific implementation manner of clustering and grouping the tasks in the initial task sequence is as follows: For the target position of each task, the shortest feasible path from the preset take-off point to the target position is searched in the flyable area of the three-dimensional dynamic environment model based on the A* algorithm; when calculating the path, the wind speed, wind direction and turbulence intensity are taken as dynamic cost factors and integrated into the path cost function, so as to preferentially select the path with lower wind speed, weaker turbulence intensity and higher safety level; at the same time, the path height is adjusted according to the atmospheric visibility data to ensure that the target is within the effective observation range; finally, the preferred path of each task is obtained; The preferred path length L of each task is calculated i The merging saving value S of any two tasks is calculated ij The improved saving value S is calculated based on the average safety level score of the preferred path of the task and the time window compatibility coefficient en ; The tasks are arranged in descending order of S en value, starting from the highest S en value, two paths are attempted to be merged, that is, the merging result is checked for constraints, if the constraints are met, the paths are merged and the task group is updated; When there is no effective merging, stop and output each task subset.

[0026] It needs to be further explained that the formula for calculating the merging saving value is: In the formula, D0 represents the initial coordinates of the unmanned aerial vehicle, dist(A, B) represents the preferred path length from the coordinates of point A to point B, T x represents the task point coordinates of the xth task; The formula for calculating S en is: In the formula, S x is the safety level score of the area where the task point is located, a ij is the time window compatibility coefficient, and the calculation formula is In the formula, t x is the time window length of task x, and t ij represents the length of the overlapping part of the time window of task i and task j. The constraint check includes: the length of the merged path L new <D max , and the task sequence after merging can be completed within the respective time windows of all tasks.

[0027] For each task subset, a vehicle routing problem model with a time window is constructed based on the current position of the unmanned aerial vehicle, the task position coordinates, the estimated flight time of the inspection path and the time window constraint; the genetic algorithm is used to solve the model to optimize the execution order of the tasks in the group, the goal is to minimize the total flight distance and time of the flight and to ensure that all tasks are completed within their time windows; According to the flyable area safety level calculation task safety index, and check whether the optimized task sequence exists time window conflict, task safety index is lower than threshold or exceeds single maximum flight distance constraint, if there is conflict, adjust the task grouping or sequence order, until all constraints are met; Further, in the above technical scheme, the calculation method of the task safety index is: ; In the formula, I is the task safety index, n is the number of paths covered in the task subset, S i is the average safety level score of the i-th path, l i is the length of the i-th path, L is the total length of the n paths, P is the task priority, β is the priority weight coefficient, △T i is the turbulence intensity change rate, γ is the meteorological mutation penalty coefficient, δ(t) is the time urgency factor, t is the remaining time of the end of the task time window; It needs to be further explained that the value range of the task safety index I is I∈[0, 201], when I<100, it is determined to be lower than the threshold, triggering the adjustment of the task grouping or sequence order; The priority weight coefficient and the meteorological mutation penalty coefficient are obtained through the mapping set of historical data and corresponding coefficients established in the database, the default value of the priority weight coefficient is β=0.7, and the default value of the meteorological mutation penalty coefficient is γ=0.4; The calculation formula of the time urgency factor δ(t) is: ; The task allocation scheme is output, which includes the number of unmanned aerial vehicle flights, the task sequence executed by each flight, the expected take-off time of each flight, the total flight distance, and the expected execution time of each task.

[0028] Task safety index calculation example: Scenario one: the average safety level scores of the task paths are S1=0.8, l1=200m, S2=0.6, l2=300m, the task priority P=0.9, the turbulence change rate △T i =0.4, t=90min, the priority weight coefficient β=0.7, and the meteorological mutation penalty coefficient γ=0.4; δ(t)=1-0.5tanh(0.05×30)=0.548; I=100×(0.8×200+0.6×300) / (200+300)×e (0.7×0.9) ×(1-0.4×1 / 7)×0.548≈38.2; Conclusion: I<100, triggering the adjustment of the task grouping or sequence order.

[0029] Scenario two: the average security level score of the task path is S1=0.95, l1=150m, S2=0.90, l2=250m, the task priority P=0.7, the turbulence change rate △T i =0.25, t=180min, the priority weight coefficient β=0.7, and the weather mutation penalty coefficient γ=0.4. Calculation process: δ(t)=1;

[0030] I=100×(0.95×150+0.9×250) / (150+250)×e (0.7×0.7) ×(1-0.4×0)×1≈150; Conclusion: I>100, and the task subset is executable.

[0031] S4, the control terminal issues a task to the unmanned aerial vehicle according to the task allocation scheme; S5, the unmanned aerial vehicle executes the task and adjusts the flight path according to real-time weather data, and the control terminal generates an inspection report after the task is completed.

[0032] Further, in the above technical solution, the adjustment of the flight path is specifically analyzed as follows: If any of the following situations is detected, the path dynamic adjustment is triggered: The turbulence intensity index of the current path segment exceeds the turbulence intensity threshold; The crosswind speed exceeds 70% of the upper limit of the wind resistance of the unmanned aerial vehicle; The atmospheric visibility decreases to cause the target observation clarity to be lower than a preset value; It should be further pointed out that the default value of the turbulence intensity threshold is 0.6.

[0033] The adjustment strategy includes: Replanning the local path based on the updated environment model to avoid high-turbulence areas and strong-wind areas; Adjusting the flight height dynamically according to the real-time visibility to ensure the observation effectiveness; If the energy consumption of the adjusted path exceeds the remaining endurance mileage, the control terminal is requested to issue a task termination or return instruction.

[0034] Further, in the above technical solution, the control terminal generates an inspection report after the task is completed, specifically including: Task execution timing log, including takeoff time, arrival time, departure time, and return time of each task; Deviation analysis of the actual flight path and the preset path; Influence statistics of weather data on flight energy consumption; Path dynamic adjustment times and reason classification; Task completion status flag, including success, partial success, failure.

Claims

1. A method for optimizing the autonomous inspection task allocation of a low-altitude unmanned aerial vehicle (UAV), characterized in that, It includes a data acquisition module, an environment modeling module, a dynamic allocation module, a drone, and a control terminal, and specifically includes the following steps: S1. Collect task attribute data, UAV status data, geographic environment data and real-time meteorological data through the data acquisition module; S2, the environmental modeling module integrates geographic environmental data and real-time meteorological data to construct a three-dimensional dynamic environmental model; S3. The dynamic allocation module generates a task allocation scheme based on the constraints of the maximum flight distance and minimum path safety threshold of the UAV in a single run, with the goal of minimizing the number of runs. S4. The control terminal issues tasks to the UAV according to the task allocation plan; S5. The drone performs the mission and adjusts its flight path based on real-time weather data. The control terminal generates an inspection report after the mission is completed.

2. The method for optimizing the autonomous inspection task allocation of a low-altitude flying UAV as described in claim 1, characterized in that: The task attribute data includes: task location coordinates, task priority, and window time; the UAV status data includes: remaining UAV range; the geographic environment data includes: digital elevation model, obstacle distribution data, no-fly zone boundary coordinates, and land cover type data; the real-time meteorological data includes: wind speed, wind direction, temperature, precipitation probability, atmospheric visibility, and turbulence intensity index.

3. The method for optimizing the autonomous inspection task allocation of a low-altitude flying UAV as described in claim 1, characterized in that: The construction process of the three-dimensional dynamic environment model includes: A static terrain skeleton is generated by overlaying digital elevation models with obstacle distribution data. Real-time meteorological data is interpolated using finite element meshes to generate spatial distribution heat maps of wind speed and turbulence intensity. By integrating the boundary coordinates of no-fly zones with land cover type data, the safety level of flyable areas is marked.

4. The method for optimizing the autonomous inspection task allocation of a low-altitude flying UAV as described in claim 3, characterized in that: The formula for calculating the security level is as follows: In the formula, S represents the safety level score, V represents the real-time wind speed, and T represents the wind speed. i C is the turbulence intensity index. t is the land cover type coefficient, w1, w2, w3 are weighting coefficients, and k is the wind speed attenuation factor.

5. The method for optimizing the autonomous inspection task allocation of a low-altitude flying UAV as described in claim 1, characterized in that: The specific analysis method for the generated task allocation scheme is as follows: Based on task priority and task time window, all tasks are sorted to generate an initial task sequence; Using the maximum single flight distance of the UAV as a constraint, an improved saving algorithm is adopted to cluster the tasks in the initial task sequence into several groups of task subsets that can be completed in a single flight. The goal is to minimize the total number of UAV flights required. For each task subset, a vehicle routing problem model with a time window is constructed based on the current position of the UAV, the coordinates of the task position, the estimated flight time of the inspection path, and the time window constraint. The genetic algorithm is used to solve the model to optimize the execution order of tasks within the group. The goal is to minimize the total flight distance and time of the sortie and ensure that all tasks are completed within their time windows. The mission safety index is calculated based on the safety level of the flyable area, and the optimized mission sequence is checked for time window conflicts, mission safety index below the threshold or exceeding the single maximum flight distance constraint. If there are conflicts, the mission grouping or sequence order is adjusted until all constraints are met. The output includes the number of drone sorties, the mission sequence for each sortie, the estimated takeoff time for each sortie, the total flight distance, and the mission allocation scheme with the estimated execution time for each mission.

6. The method for optimizing the autonomous inspection task allocation of a low-altitude flying UAV as described in claim 5, characterized in that: The task safety index is calculated as follows: In the formula, I represents the task safety index, n represents the number of paths covered in the task subset, and S... i To score the security level of the i-th path segment, l i Let L be the length of the i-th path segment, L be the total length of the n paths, P be the task priority, β be the priority weight coefficient, and ΔT be the path length. i denoted as the rate of change of turbulence intensity, γ as the meteorological sudden change penalty coefficient, and δ(t) as the time urgency factor.

7. The method for optimizing the autonomous inspection task allocation of a low-altitude flying UAV as described in claim 1, characterized in that: The specific analysis method for adjusting the flight path is as follows: If any of the following conditions are detected, dynamic path adjustment will be triggered: The turbulence intensity index of the current path segment exceeds the turbulence intensity threshold; The crosswind speed exceeded 70% of the drone's wind resistance limit; Decreased atmospheric visibility resulted in target observation clarity falling below the preset value; The adjustment strategies include: Local paths are replanned based on updated environmental models to avoid areas of high turbulence and strong winds; The flight altitude is dynamically adjusted based on real-time visibility to ensure the effectiveness of observations; If the energy consumption of the adjusted route exceeds the remaining driving range, a mission abort or return command will be requested from the control terminal.

8. The method for optimizing the autonomous inspection task allocation of a low-altitude flying UAV as described in claim 1, characterized in that: The control terminal generates an inspection report upon completion of the task, specifically including: Task execution sequence log; Deviation analysis between actual flight path and preset path; Statistics on the impact of meteorological data on flight energy consumption; Number of times the path was dynamically adjusted and the classification of the reasons; Task completion status marker.

Images